The NTP conducts assays using growth, development, reproduction and movement of the nematode, C. elegans. We collaborated with the C. elegans laboratory to develop new statistical models that describe normal growth and assess effects of chemical exposure on growth. Rodent feed and bedding may contain compounds that affect the same endpoints that are the target of NTP's testing program, such as reproduction and development. For example, phytoestrogens in feed may have estrogenic effects of the same strength as a chemical being tested, or bedding may contain mold byproducts that can mask responses to exposure to asthma-causing agents. We are working with the Quality Assurance Laboratory to quantify these effects and develop guidelines for limits on contaminants in feed and bedding of rodent studies. The NTP conducts rodent micronucleus tests as part of its battery of genetic toxicity testing. Normally, as red blood cells mature, they shed their nucleus. If chromosome damage has occurred in the red blood cells, small parts of the nucleus (micronuclei) may remain in the cell. In the micronucleus test, blood or bone marrow of mice or rats exposed to a chemical is examined under the microscope and numbers of mature red blood cells containing micronuclei are counted. Recently, a new technique using flow cytometry has been proposed for counting micronuclei in a less mature population of red blood cells. An advantage of this technique is that many more cells can be examined per animal than is feasible with microscopy. We contributed to studies comparing microscopic and flow cytometry enumerations of micronuclei. Blood and bone marrow samples from studies of nine chemicals in mice and rats were evaluated using both microscopic and flow cytometry techniques. We found that the two techniques produced very similar micronucleus counts and they resulted in the same conclusion about whether a chemical is genotoxic. In the mid-1990s, several NTP studies of chemicals were partially compromised by an infection of Heliobacter hepaticus. This bacterium causes hepatitis in mice which often leads to liver tumors. Because liver tumors were observed in these studies, it was not clear whether H. hepaticus or the chemical exposure was responsible. For most of these studies, the chemical caused cancer in other organs unaffected by the infection, so the carcinogenic effect of the chemical could be determined. For triethanolamine, however, tumors were observed only in the liver. Several years later, the NTP conducted a second, identical study of triethanolamine in mice that were free of H. hepaticus infection. However, at that time, the diet that the mice were fed had been changed and animals were larger in the second study. Because liver tumor rates increase with body weight, among other factors, our analyses included statistical modeling of liver tumor rates using historical control data from the each study. We analyzed data from both studies and determined that liver tumor rates in males were only slightly increased by triethanolamineexposure and that the H. hepaticus infection accounted for the statistically significant increase seen in the initial study. Liver tumors in female mice were significantly increased by triethanolamine exposure, independently of the H. hepaticus infection. Because study design is critical to the collection of high quality data and may control factors involved in extraneous variability of data, we contributed to NTP study design teams for 3 chemicals. We also provided statistical design and analysis advice to the NTP's Host Susceptibility Branch and Biomolecular Screening Branch as they seek to develop new methods for evaluating the effects of environmental exposures on rodents and, ultimately, humans.